The Problems with Measuring and Using Happiness for Policy Purposes Mark D

The Problems with Measuring and Using Happiness for Policy Purposes Mark D. White

MERCATUS RESEARCH Mark D. White. “The Problems with Measuring and Using Happiness for Policy Purposes.” Mercatus Research, Mercatus Center at George Mason University, Arlington, VA, July 2015.

ABSTRACT Many governments around the world are considering measures of happiness or subjective well-being as alternatives to gross domestic product (GDP) for the purpose of guiding economic policymaking. Compared to GDP, happiness measures promise to better capture the quality of life of a nation’s citizens and lead to policies that are more effective and equitable. However, there are a number of problems with the concept of happiness that policymakers should be aware of before adopting it as a policy tool. In this paper, I focus on three interrelated aspects of happiness—definition, measurement, and policy implementation—and explain why each renders happiness a poor guide for policy. In general, happiness is a vague, multifaceted, and subjective phenomenon that is difficult to define precisely enough for measurement, hard to measure in a way that allows meaningful comparison between individuals and groups, and fraught with ethical complexities that complicate policy implementation.

JEL codes: D6, D63, E6, E61, H1, I31

Keywords: economic policy, GDP, happiness, well-being, welfare, measurement, psychology, philosophy

Release: July 2015 Originally released as a Mercatus Working Paper, December 2014.

The opinions expressed in Mercatus Research are the author’s and do not represent official positions of the Mercatus Center or George Mason University. conomists have long expressed dissatisfaction with standard national-accounting figures such as gross domestic product (GDP). In fact, the shortcomings of GDP are so widely accepted that every introductory macroeconomics student learns them. For Eexample, GDP misses black-market transactions and nonmarket production such as household labor, both of which contribute to national output. Thus, GDP underestimates the actual value of the total work performed by a nation’s citizens and specifically underestimates types of work usually done by certain kinds of citizens (for instance, it neglects the unpaid labor of stay-at-home partners, who are disproportionately women). Also, GDP doesn’t account for the value of byproducts of production, including negative externalities such as pollution as well as positive externalities such as neighborhood improvement. Most generally, GDP does not measure the true quality of life of citizens and residents but is, at best, an imperfect measure of the resources available to the average person to pursue his or her goals (and even this assumes a certain level of wealth equality that GDP does not account for). It is the last of these objections that, along with developments in experi- mental psychology and behavioral research, has motivated discussion among economists and policymakers about measuring well-being directly, using measures of happiness or subjective well-being developed by psychologists, rather than relying on a flawed and incomplete economic proxy. As Richard Layard writes, “happiness should become the goal of policy, and the progress of national happiness should be measured and analyzed as closely as the growth of [gross national product].”1 First implemented under the banner of “gross domestic happiness” by King Jigme Singye Wangchuck of Bhutan in the early 1970s, the idea of basing policymaking on measurements of happiness and well-being (as well as other factors, including GDP) was given renewed life in 2009 by French president Nicolas Sarkozy, who enlisted renowned economists Joseph

1. Richard Layard, Happiness: Lessons from a New Science (New York: Penguin, 2005), 147.

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 3 Stiglitz, Amartya Sen, and Jean-Paul Fitoussi to study its feasibility.2 The United Nations passed a resolution in 2011 encouraging member states to track aggregate happiness and published the first World Happiness Report in 2012. This was followed by the Organisation for Economic Co-operation and Development’s Guidelines on Measuring Subjective Well-Being in 2013.3 Finally, in recent years the Office of Management and Budget in the United States has “There are recommended in its annual reports to Congress that alter- numerous native measures of well-being be explored as a supplement to using GDP and benefit-cost analysis for the purposes of problems with policymaking and regulation.4 As data collection and com- the concept of puting power continue to expand, it is safe to assume that the appeal of happiness-based policy will grow as well. happiness or Given the significant shortcomings of GDP as an esti- subjective well- mate of national well-being, policymaking based on direct being that cast measures of happiness has obvious appeal as a way to reori- ent government action toward having a more direct and doubt on its positive impact on the lives of a nation’s citizens. However, usefulness as a there are numerous problems with the concept of happiness or subjective well-being that cast doubt on its use- policymaking fulness as a policymaking tool. In this paper, I will survey tool.” problems in three areas: definition, measurement, and implementation. In short, (1) happiness is too vague and multifaceted a concept to define clearly and precisely; (2) even if it could be defined, it is an essentially qualitative concept that is difficult to quantify and measure with any confidence; and (3) even if it could be measured, designing

2. Joseph E. Stiglitz, Amartya Sen, and Jean-Paul Fitoussi, Mismeasuring Our Lives: Why GDP Doesn’t Add Up (New York: New Press, 2010), http:// www.stiglitz-sen-fitoussi.fr/en/index.htm. 3. J. F. Helliwell, R. Layard, and J. Sachs, eds., World Happiness Report (New York: Earth Institute, 2012), http://issuu.com/earthinstitute/docs /world-happiness-report; OECD Guidelines on Measuring Subjective Well- Being (Paris: Organisation for Economic Co-operation and Development, 2013), http://www.oecd.org/statistics/guidelines-on-measuring-subjective -well-being.htm. 4. These reports can be found at Office of Management and Budget, “OIRA Reports to Congress,” http://www.whitehouse.gov/omb/inforeg_regpol _reports_congress; the most recent two reports (from 2013 and 2014) have downplayed this idea significantly.

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 4 and implementing policy based on happiness raises a number of ethical and political issues that cannot be solved by improving the science of measurement. In the end, this paper sounds a strong cautionary note to policymakers who might consider incorporating measures of happiness or subjective well-being into their policymaking processes and deliberations.

DEFINITION The most fundamental problem with happiness-based policymaking is that “happiness” is a notoriously difficult concept to define. Like “justice” or “beauty,” happiness is a vague term that means different things to different people; as a consequence, even though everyone knows what it means in various situations, we would be hard-pressed to come up with a single definition that captures all those aspects for every person. Philosopher Sissela Bok devotes an entire chapter of her book Exploring Happiness to this issue; the title of the chapter, “Discordant Definitions,” is indicative of her view.5 As she writes, “such abstract terms provide ideal vessels into which people can pour quite different, sometimes clashing, meanings.”6 Economists and policymakers rely on the definitions proposed by philosophers and psychologists, but, as we shall see, they are of little help. Even though philosophers have studied happiness for thousands of years, they have failed to arrive at a single, canonical definition. Aristotle wrote elo- quently about the concept of eudaemonia, which describes a life of virtuous flourishing and realized excellence in one’s chosen pursuits. This idea resem- bles modern psychologists’ notion of life satisfaction but is ethically much richer.7 In his seminal work on utilitarianism, Jeremy Bentham relied on a far simpler version of happiness, a hedonic conception based on pleasure versus pain. John Stuart Mill later elaborated on this, positing higher and lower forms of pleasure as reflected in his famous statement that “it is better to be a human being dissatisfied than a pig satisfied; better to be Socrates dissatisfied than a fool satisfied.”8 Modern philosophers accept all these conceptions as valid

5. Sissela Bok, Exploring Happiness: From Aristotle to Brain Science (New Haven, CT: Yale University Press, 2010), chapter 3. 6. Ibid., 57. 7. On Aristotle’s conception of eudaemonia, see Stanford Encyclopedia of Philosophy, s.v. “Aristotle’s Ethics,” by Richard Kraut, revised April 21, 2014, http://plato.stanford.edu/entries/aristotle-ethics/. 8. Jeremy Bentham, An Introduction to the Principles of Morals and Legislation (Oxford: Clarendon Press, 1786; 1823), http://www.econlib.org/library/Bentham/bnthPMLCover.html; John Stuart Mill, Utilitarianism (London: Longmans, Green, and Co., 1879), http://www.gutenberg.org/ebooks/11224. (The specific quote in the text is from chapter 2.) Philosopher Martha Nussbaum locates Mill’s version

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 5 interpretations of happiness but no one of them as definitive; as Wayne Sumner writes, “no simple theory about the nature of happiness enjoys much support among philosophers; there is not even agreement that such a theory is possible. About the only thing everyone agrees on is that happiness is a complex and multi-faceted notion, one not easily reduced to a formula or slogan.”9 Psychologists who study happiness—often known as positive psychologists, in contrast to the majority who study mental disorders such as depression and anxiety—also have not settled on a single definition. Ed Diener, one of the founding figures in the area, wrote with several colleagues that “the nature of happiness has not been defined in a uniform way. Happiness can mean pleasure, life satisfaction, positive emotions, a meaningful life, or a feeling of contentment, among other concepts.”10 Mirroring Sissela Bok’s philosophi- cal viewpoint, psychologist Daniel Gilbert writes that happiness “is nothing more or less than a word that we word makers can use to indicate whatever we please. The problem is that people seem pleased to use this one word to indicate a host of different things, which has created a tremendous terminological mess.”11 Many psychologists prefer to use the term subjective well-being, coined by Diener, to encompass both hedonic pleasure in the moment and life satisfaction over time, but the next section will show the additional problems this composite definition causes for measurement. More than anybody else, philosopher Daniel Haybron has examined the various conceptions of happiness used by both philosophers and psychologists. Haybron categorizes these definitions into three types, including the two already mentioned: hedonic pleasure and life satisfaction.12 Although the nature of each is very different, these two conceptions share an emphasis on evaluation: each demands that a person assess his or her happiness either at a point in time or over a lifetime. Haybron, on the other hand, favors a third meaning of happiness: a person’s emotional state itself as opposed to his or

of happiness between Bentham’s simple hedonism and Aristotle’s rich eudaemonism. See Nussbaum, “Mill between Aristotle and Bentham,” in Economics & Happiness: Framing the Analysis, ed. Luigino Bruni and Pier Luigi Porta (Oxford: Oxford University Press, 2005). Nussbaum also criticizes psychologists’ practice of linking their conceptions of happiness to those of Bentham and Aristotle. See Nussbaum, “Who Is the Happy Warrior? Philosophy Poses Questions to Psychology,” in Law & Happiness, ed. Eric A. Posner and Cass R. Sunstein (Chicago: University of Chicago Press, 2010). 9. L. W. Sumner, Welfare, Happiness, and Ethics (Oxford: Oxford University Press, 1996), 139. 10. Ed Diener, Christie Napa Scollon, and Richard E. Lucas, “The Evolving Concept of Subjective Well-Being: The Multifaceted Nature of Happiness,” in Assessing Well-Being: The Collected Work of Ed Diener, ed. Ed Diener (Dordrecht: Springer, 2009), 68. 11. Daniel Gilbert, Stumbling on Happiness (New York: Vintage, 2005), 33. 12. Daniel Haybron, The Pursuit of Unhappiness: The Elusive Psychology of Well-Being (Oxford: Oxford University Press, 2008).

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 6 her cognitive impression of it. Focusing on evaluation puts the cart before the horse; as economists Marc Fleurbaey and Didier Blanchet write about life satisfaction, “it is not satisfaction that makes a good life, but a good life that gives satisfaction.”13 Even those who urge the measurement of happiness for policy purposes recognize the feeling behind the assessment: Richard Layard writes that “happiness is feeling good,” and Ed Diener, with his son Robert Biswas- Diener, emphasizes feeling as well as evaluation: “Happiness is the name we put on thinking and feeling positively about one’s life.”14 Each of these three conceptions of happiness is familiar, unique, and appropriate in particular circumstances—but none is comprehensive or exhaustive, and therein lies the difficulty. In the end, happiness may function as an “umbrella” term, covering a number of related phenomena without speci- fying any one of them. In the spirit of Judge Potter Stewart’s famous words about obscenity, we know happiness when we see it (or experience it). But some researchers take this too far. For example, leading positive psychologist Sonja Lyubomirsky writes that “most of us don’t need a definition of happiness because we instinctively know whether we are happy or not,” and Daniel Gilbert calls happiness “the you-know-what-I-mean feeling.”15 Nevertheless, a widely accepted definition is necessary to generate reasonable confidence in the meaning and relevance of the results of any attempt at measurement, even if only to ensure that the respondents to a survey have the same understanding of happiness that the researchers had in mind when they designed it. Consider many happiness surveys, which contain questions such as “how happy do you feel at the present moment” or “how well do you feel your life is going,” with responses solicited on a numerical scale with labels such as “very happy,” “moderately unhappy,” or “neutral.” For example, the World Values Survey asks,

All things considered, how satisfied are you with your life as a whole these days? Using this card on which 1 means you are

13. Marc Fleurbaey and Didier Blanchet, Beyond GDP: Measuring Welfare and Assessing Sustainability (Oxford: Oxford University Press, 2013), 171. In the technical language of economists, they write, “It is apparently easy for economists to forget that when the economic model makes the individual maximize u(x), this means that the individual cares about x, not u(x). If individuals cared about u(x), they would spend their time working on their mind-set rather than changing the world around them.” Ibid., 202. 14. Layard, Happiness, 6; Ed Diener and Robert Biswas-Diener, Happiness: Unlocking the Secrets of Psychological Wealth (Malden, MA: Blackwell, 2008), 4. 15. Sonja Lyubomirsky, The How of Happiness: A New Approach to Getting the Life You Want (New York: Penguin, 2007), 32; Gilbert, Stumbling on Happiness, 35.

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 7 “completely dissatisfied” and 10 means you are “completely satisfied” where would you put your satisfaction with life as a whole?16

One problem with such questions is that they leave the meaning of happiness up to the respondent, who may have a very different conception of it than the researcher has. In one study, psychologists Luo Lu and Robin Gilmour asked students in the United States and China to explain what happiness meant. While the American students’ answers emphasized individualism and materi- alism, Chinese students’ answers were based on communitarian and spiritual concerns—and their answers on happiness surveys would reflect these different conceptions.17 Everyone may “know” what happiness is, but that doesn’t mean that everyone shares the same idea of what it means (or in what context they should assess it, as I will discuss below). Researchers need a clear, precise definition of happiness in order to know what to measure and how to interpret their results. But such a definition does not and cannot exist, because of the vague and multifaceted nature of happiness itself. Even if there were a useful, canonical definition of happiness, however, there are other serious problems with trying to measure an essentially qualitative concept in quantitative terms, which I will survey in the next section.

MEASUREMENT From this point on, I will assume for the sake of argument that survey respondents have a clear understanding of what the researcher’s conception of happiness is. This still leaves a number of questions regarding the scale used to measure happiness and how it is to be understood. First, there are issues with the extreme ends of the scale, both of which are difficult to interpret in a way that would allow the results to be normalized, combined, and averaged. Take the lower bound on any scale of happiness: does this mean an absence of happiness, or does it indicate profound unhappiness? In other words, does assessed or felt happiness have a meaningful zero point representing abject misery, or should zero be the point separating happiness from unhappiness (which should be represented by negative numbers)? The upper bound faces a similar dilemma: what is “maximal” happiness supposed to represent? If respondents

16. OECD Guidelines, 249. 17. Luo Lo and Robin Gilmour, “Culture and Conceptions of Happiness: Individual Oriented and Social Oriented SWB,” Journal of Happiness Studies 5 (2004): 269–91.

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 8 answer a survey question about happiness with the highest value on the scale, does this imply that they literally cannot imagine being any happier, or merely that they are as happy as they can be in the limited context of their current life circumstances? To help make these questions concrete, consider “Cantril’s Ladder of Life,” included in the Gallup World Poll, which takes the following form:

Please imagine a ladder with steps numbered from zero at the bottom to ten at the top. Suppose we say that the top of the ladder represents the best possible life for you and the bottom of the ladder represents the worst possible life for you. If the top step is 10 and the bottom step is 0, on which step of the ladder do you feel you personally stand at the present time?18

The concepts of “best possible life for you” and “worst possible life for you” are designed to be personal and subjective, which would be of some value for tracking one person’s level of happiness over time, but are too sensitive to context and circumstance to standardize the scale for comparison between people. (Even one person may change his or her impression of the best and worst possible life over time, compromising that value as well.) In her work on the economics of happiness, Carol Graham writes of “happy peasants and frustrated achievers.” As she explains, peasants may report tremendous happiness despite meager circumstances, either because they have adapted to them or because they literally cannot imagine any better “possible life.”19 Similarly, the achievers may report low happiness despite great material and emotional success because they have acclimated to their fortune and constantly focus on the next level; their “best possible life” is unbounded. This is hardly a theoretical or hypothetical problem: many researchers have found comparable levels of reported happiness in first-world and third-world countries despite extremely disparate conditions and opportunities. This is an aspect of the Easterlin paradox regarding the disconnect between wealth and happiness, which spawned happiness research among economists (as well as

18. See “Understanding How Gallup Uses the Cantril Scale,” Gallup website, accessed November 25, 2014, http://www.gallup.com/poll/122453/understanding-gallup-uses-cantril-scale.aspx. On Cantril’s Ladder, see Hadley Cantril, The Pattern of Human Concerns (New Brunswick, NJ: Rutgers University Press, 1965). 19. Carol Graham, Happiness around the World: The Paradox of Happy Peasants and Miserable Millionaires (Oxford: Oxford University Press, 2010), and The Pursuit of Happiness: An Economy of Well-Being (Washington, DC: Brookings Institution, 2011).

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 9 work on adaptation by philosophers and psychologists).20 Therefore, even if respondents in different socioeconomic groups share the same conception of happiness as researchers do, they may have very different interpretations of what “very happy” and “mildly unhappy” mean, in which each person under- stands these categories within his or her own unique life experience. Combined with the ambiguity and dependence on context of the terms themselves, the varied and unpredictable interpretations of happiness surveys across cultures and continents make any meaningful comparison between them difficult. There is also the problem of reading too much precision into a vaguely defined numerical scale applied to qualitative phenomena. Again, for the sake of argument we can assume that all survey respondents understand the head- ings on a survey’s numerical scale the same way. Nonetheless, transferring the qualitative categories of “more” and “less” happiness into that numerical scale implies a regularity that does not exist. Let’s say that, on a given scale, three represents “mildly happy,” four represents “moderately happy,” and five represents “very happy.” Translating these categories into a numerical scale and performing almost any operation with them, even simply averaging them, assumes that the units are equal: just as the difference between three and four is equal to the difference between four and five, we are to believe that the difference between “mildly happy” and “moderately happy” is equivalent to the difference between “moderately happy” and “very happy.” But we have no reason to believe this in the case of even one person, much less many people. As economist and philosopher Deirdre McCloskey writes,

If you like the temperature in Chicago today better than the one on January 15, you might be induced by the interviewer to assign 2.76 to today and a 1.45 to January 15. But such an assign- ment is of course arbitrary. . . . By contrast, an interval scale, such as Fahrenheit or Celsius temperature on the two days in question, does measure, 1-2-3. . . . We couldn’t base a physics on asking people whether today was “hot, nice, or cold” and expect to get anything quantitative out of it.21

20. On the Easterlin paradox, see Richard Easterlin, “Does Economic Growth Improve the Human Lot? Some Empirical Evidence,” in Nations and Households in Economic Growth: Essays in Honor of Moses Abramowitz, ed. Paul A. David and Melvin W. Reder (New York: Academic, 1974). For a skep- tical reexamination, see Betsey Stevenson and Justin Wolfers, “Economic Growth and Subjective Well-Being: Reassessing the Easterlin Paradox,” Brookings Papers on Economic Activity 39, no. 1 (2008): 1–87. 21. Deirdre McCloskey, “Happyism,” New Republic, June 8, 2012, http://www.newrepublic.com /article/politics/magazine/103952/happyism-deirdre-mccloskey-economics-happiness.

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 10 Assigning fixed quantitative values to arbitrary qualitative categories results only in false precision and gives the impression that the measurement contains more information than the researcher actually has. Some researchers compound this error by combining values from the same numerical scale for different types of happiness. In her Subjective Happiness Scale, Sonja Lyubomirsky combines the two conceptions of evaluated “Assigning fixed happiness in one survey, which consists of four questions covering how happy you feel and how your happiness com- quantitative pares to that of others (emotional state) and how happy and values to arbitrary unhappy you are with your life (satisfaction), each of which is to be answered on a scale from zero to seven.22 Then she qualitative asks the respondents to add up the values from the four categories questions to compute their overall happiness measure. This results only in compounds the problem identified above: not only does it assume that the intervals between numerical answers to false precision any given question are equivalent, but it also assumes that and gives the the values and intervals are equivalent between questions. According to her framework, a person who reports threes impression that for emotional state and sixes for life satisfaction—such as the measurement a relatively successful person going through a momentary rough patch—is seen as exactly as happy as a person who contains more reports sevens for emotional state and twos for life satisfac- information than tion—such as a person who enjoys a small success while still the researcher judging himself to be on the wrong path. These two results will report the same overall score even though the nature of actually has.” each person’s happiness is very different, both qualitatively and quantitatively.23 By comparison, the Organisation for Economic Co-operation and Development’s Guidelines on Measuring Subjective Well-Being (Annex B) constructs composite measures of happiness based on questions from

22. Lyubomirsky, How of Happiness, 33. 23. Ironically, the possibility of a discrepancy between reports of momentary happiness and life satisfaction should be seen as a strength, not a flaw, of multidimensional happiness studies: since the two types of happiness are distinct, a researcher should be interested in each of them independently of the other. If they were closely correlated, there would be no benefit from measuring both of them, but because they can diverge, they should not be conflated by summing their scores.

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 11 various different surveys covering life satisfaction, pleasure, and feelings, each of which is averaged between people—which is problematic on its own. But the measures of each of these different areas of happiness are not summed up, lessening though not eliminating the problems with its happiness survey data.24 As well as providing another example of the problem with quantifying happiness, the possibility of such discrepancies illustrates the wide array of attitudes toward life that confound reported happiness measures, especially those concerning life satisfaction, which many regard as more important than passing emotional states. Philosopher Ronald Dworkin writes of Jack and Jill, both of whom are “reasonably successful in their chosen occupations” but each of whom sees the value of a life, as well as the “best possible life for you,” very differently from the other:

They take roughly the same enjoyment from their day-to-day life. But Jack (who has been influenced by genre painting) thinks that any ordinary life fully engaged in projects is a life of value, while Jill (perhaps because she has taken Nietzsche to heart) is much more demanding. Jack thinks, for example, that the life of a busy peasant who achieves very little and leaves nothing behind is full of value, while Jill thinks that such a life is only full of failure. If each is asked to assess the overall value of his or her own life, Jack would rate his high and Jill hers low.25

Presumably, Jack and Jill would interpret the endpoints and labels on a numerical scale of life satisfaction very differently based on their disparate opinions about what it means for a life to have value. Furthermore, personal circumstances can cause a person to adapt to certain levels of happiness and life satisfaction (as in Graham’s “happy peasants and frustrated achievers”). This may not be of great concern with Jack and

24. The criticisms of measuring happiness summarized herein should not be extrapolated to surveys in general, which have valuable uses in social-scientific research and policy. For instance, polls that ask people to rate their opinion of a politician or proposed legislation on a scale of 1 to 10 invoke an explicitly numerical response to a simple question which is relatively clear, and the results are understood merely to be an indication of voter sentiment, not to be used as a target for policymaking. The problems with measuring happiness derive from the vague and qualitative nature of the phenomenon itself and from the problems it introduces into the interpretation of aggregate survey results as a guide to national economic policymaking. 25. Ronald M. Dworkin, Sovereign Virtue: The Theory and Practice of Equality (Cambridge, MA: Harvard University Press, 2000), 36; see chapter 1 in general for concerns about scales and rankings of well-being.

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 12 Jill, who are both living successful lives in a first-world country with the freedom to pursue the interests they choose, but in less affluent areas of the world, adaptation can mask severe deprivation with the appearance of happiness. As Amartya Sen writes,

The hopeless beggar, the precarious landless laborer, the dominated housewife, the hardened unemployed or the over- exhausted coolie may all take pleasures in small mercies, and manage to suppress intense suffering for the necessity of continuing survival, but it would be ethically deeply mistaken to attach a correspondingly small value to the loss of their well- being because of this survival strategy.26

McCloskey writes in a similar vein:

If a man tormented by starvation and civil war in South Sudan declares that he is “happy, no, very happy, a regular three, mind you,” we have learned something about the human spirit and its sometimes stirring, sometimes discouraging, oddity. But we inch toward madness if we go beyond people’s lips and claim to read objectively, or subjectively, their hearts in a 1-2-3 way that is comparable with their neighbors or comparable with the very same South Sudanese man when he wins an immigration lottery and gets to Albany.27

As she notes, this phenomenon displays admirable perseverance in the face of horrific circumstances, but it bears little relation to well-being as researchers wish to measure it, nor to what policymakers want to increase. We should respect people’s unique conceptions and judgments of their own happiness while acknowledging the difficulties with making comparisons between them and basing policy decisions upon them. Indeed, were we to take such reports as valid and implement policy based on them, governments would end up helping the dissatisfied affluent rather than the truly needy (as described in the next section). In the end, happiness is an inherently qualitative concept that cannot be translated into quantitative terms with any reliable degree of precision or

26. Amartya Sen, On Ethics and Economics (Oxford: Blackwell, 1987), 45–46. 27. McCloskey, “Happyism.”

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 13 comparability between people, reflecting “the gap between the open-ended nature of many dimensions of life and the bounded scale imposed by questionnaires.”28 After acknowledging many of the problems detailed here, psychologist Daniel Gilbert admits that

when we say with moderate precision what we mean by words such as happiness, we still can’t be sure that two people who claim to be happy are having the same experience, or that our current experience of happiness “many is really different from our past experience psychologists of happiness, or that we are having an expe- 29 question whether rience of happiness at all. self-reported Furthermore, many psychologists question whether happiness actually self-reported happiness actually reports happiness at all. Jerome Kagan questions why “psychologists are willing to reports happiness believe that a verbal report of lifetime happiness, without at all.” any other information, is an accurate measure of a psychological state whose definition remains fuzzy.”30 Angus Campbell writes that reliance on happiness measures “is based on the assumption that all the countless experiences people go through from day to day add to . . . global feelings of well-being, that these feelings remain relatively constant over extended periods, and that people can describe them with candor and accuracy.”31 Finally, psychologists and philosophers have demon- strated that we are very poor at assessing our own emotional states in the present or predicting them in the future. For example, Daniel Haybron outlines numerous cognitive flaws that hamper our internal assessment of past and present happiness, which are compounded when we attempt

28. Fleurbaey and Blanchet, Beyond GDP, 202. 29. Gilbert, Stumbling on Happiness, 70. 30. Jerome Kagan, Psychology’s Ghosts: The Crisis in the Profession and the Way Back (New Haven, CT: Yale University Press, 2012), xix. 31. Angus Campbell, The Sense of Well-Being in America (New York: McGraw-Hill, 1981), 23.

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 14 to predict future happiness (a difficulty that he terms “affective ignorance”).32 Even if we could make accurate self-assessments, many people around the world do not have the luxury of reflecting on their emotional states or their degree of life satisfaction; as Kagan writes, “many individuals have not asked themselves how satisfied they are with their lives, and psychologists should also be uncertain of the meaning of these answers,” especially in an artificial survey environment in which the respondents’ current circumstances and context are unknown to the researcher.33 This adds yet another layer of ambiguity to a measurement process that is already fraught with problems.

POLICY IMPLEMENTATION As I consider the third challenge to happiness research, let us assume for the sake of argument that the previous two challenges have been resolved—that is, that happiness can be both defined and measured with precision and accuracy. In this section, I will focus only on the problems inherent in implementing policies based on those measures. I will show that many of these problems are familiar, arising when any single economic measure (such as GDP) is used for policy purposes, but some are unique to happiness. Together, these problems should instill researchers and policymakers with a strong sense of caution when they consider incorporating measures of happiness into their decision- making processes. I will start with the goals of happiness-based policy, which are difficult to conceptualize. Targets for GDP growth are based on factors such as population growth, technological progress, and the effects of other government policies, all of which narrow down the reasonable expectations for growth. But happiness would not seem to be limited in the same way by material concerns. Would we then want to maximize happiness, or combine it with other goals? Many scholars, such as Richard Layard and those on Nicolas Sarkozy’s team, recommend including happiness alongside more traditional measures such as GDP in a composite measure.34 But if these other measures account for aspects of the economy that are valued for their contribution to happiness, what is the argument for including them if the final target itself, happiness, is measurable?

32. Haybron, Pursuit of Unhappiness, chapter 10. See also Gilbert, Stumbling on Happiness, begin- ning with chapter 4; and George Loewenstein and David Schkade, “Wouldn’t It Be Nice? Predicting Future Feelings,” in Well-Being: The Foundations of Hedonic Psychology, ed. Daniel Kahneman, Ed Diener, and Norbert Schwarz (New York: Russell Sage Foundation, 1999). 33. Kagan, Psychology’s Ghosts, 83–84. 34. See the Layard quotation above and Stiglitz, Sen, and Fitoussi, Mismeasuring Our Lives, xxvi.

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 15 Even if we decide to focus on happiness alone, however, maximizing it may not be as simple as it would seem. For instance, as with any policy goal, maximizing happiness involves costs, which in this case would be measured in terms of for- gone happiness in order to provide a consistent basis for comparison. If, after a point, additional happiness comes only at increasing cost, there may be an “optimal” amount of happiness to target (much like sustained levels of growth or unemployment) rather than aiming always for more happiness regardless of the costs.35 In itself, this is not an argument against optimizing happiness, but simply an example of the complexity of the issue. Often, however, policymakers will have to assess tradeoffs, not simply within one common type of happiness, but between two or more qualitatively different types. Many people sacrifice short-term, experienced pleasure now for long-term, deeper fulfillment later, including going to medical school to become a doctor or forgoing desserts to live a longer, healthier life. But many measures of happiness would register this as a decline in happiness now with no way of discerning that it is an investment in yet-unrealized future happiness (of a different kind). Policymakers would have to determine how to account for such voluntary individual tradeoffs in order to avoid seeing such choices as a decline in overall well-being based on their negative impact on current pleasure. Furthermore, while growth in overall happiness may not be constrained by scarcity in the same way that GDP is, it is limited psychologically: individuals and nations can always get wealthier, but they cannot perpetually become happier. One reason is related to the way we measure happiness itself, on a finite scale with an upper bound. As McCloskey writes, “Wait a minute. The scale is 1-2-3. Of course it levels off: The ceiling, namely 3, is built into the question, and so the survey researcher gets back what she put in.”36 A more substantive reason is adaptation, the process described above by which people unconsciously adjust their expectations and reactions to increasing wealth. This aspect of adaptation is illustrated with historical perspective by economists Fleurbaey and Blanchet:

Our species has considerably evolved since it left its African cra- dle 70,000 years ago. It has progressively conquered the whole

35. On the possibility of optimal happiness, see Shigehiro Oishi, Ed Diener, and Richard Lucas, “The Optimum Level of Well-Being: Can People Be Too Happy?,” Perspectives in Psychological Science 2 (2007): 346–360; and Shigehiro Oishi and Minkyung Koo, “Two New Questions about Happiness: ‘Is Happiness Good?’ and ‘Is Happiness Better?,’” in The Science of Subjective Well-Being, ed. Michael Eid and Randy J. Larsen (New York: The Guilford Press, 2008). 36. McCloskey, “Happyism.”

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 16 planet, assimilating or eliminating other human species, and has now gained an amazing and even self-endangering power over its surroundings. It has proliferated considerably, and about two-thirds of it has reached an unprecedented level of longev- ity, consumption, and sophistication. If happiness surveys had been conducted over this whole period, one would perhaps find a remarkable stability in happiness scores over the millennia.37

A final reason is provided by psychologists in the form of set point theory, the idea that each person’s level of happiness is largely determined genetically, with small variation possible in response to temporary life circumstances.38 A related phenomenon is the hedonic treadmill, by which a person works hard to achieve a certain level of happiness only to adapt to that improved situation over time, which forces him or her to work even harder to get another boost in happiness.39 This implies that happiness-based policy is futile; as Richard Easterlin writes, “not only is public policy [focused on happiness] likely to be ineffective but there is little an individual can do to improve his or her well- being, except, perhaps, consult a psychologist.”40 Psychologists agree as well: Ed Diener and Richard E. Lucas write that “the influence of genetics and personality suggests a limit on the degree to which policy can increase [subjective well-being].”41 There are also questions regarding the sources of happiness to be used for policy deliberations, some of which may be regarded as inauthentic, askew, or even unethical. While set point theory implies that our level of happiness is largely determined by genetics, the rest can be manipulated by various means. Robert Nozick’s famous “experience machine” is a thought experiment that prompts us to imagine a machine that would stimulate our brain to make us

37. Fleurbaey and Blanchet, Beyond GDP, 163. 38. On set point theory, see Martin Seligman, Authentic Happiness: Using the New Positive Psychology to Realize Your Potential for Lasting Fulfillment (New York: Simon and Schuster, 2002), chapter 4. 39. See Shane Frederick and George Loewenstein, “Hedonic Adaptation,” in Well-Being, ed. Kahneman, Diener, and Schwarz. There is a clear analogy to repeated efforts to boost the rate of economic growth above its long-run potential using expansionary policy, in which the growth rate snaps back to normal over time but prices rise—and keep rising with continued action. 40. Richard Easterlin, “Building a Better Theory of Well-Being,” in Economics & Happiness, ed. Bruni and Porta, 30. In Beyond GDP, Fleurbaey and Blanchet note the irony in that “the ‘paradox’ in Easterlin’s paradox is that it has served to popularize the happiness challenge to the economic approach, whereas it is the best proof that the happiness approach is problematic” (p. 173). 41. Ed Diener and Richard E. Lucas, “Personality and Well-Being,” in Well-Being, ed. Kahneman, Diener, and Schwarz, 227; see also Richard E. Lucas, “Personality and Subjective Well-Being,” in Science of Subjective Well-Being, ed. Eid and Larsen.

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 17 feel as if we experienced our every desire, and then asks us whether we would prefer that to a life with agency. The most common answer is that, as pleasant as such a simulation might be, there is a greater value to “really” living and experiencing life as an active participant. Even though there is no experience machine yet, there are less fantastical ways of boosting mood, such as antide- pressants. In his book Artificial Happiness, philosopher and medical doctor Ronald W. Dworkin describes people who felt happy under medication while the abhorrent conditions in their lives continued; they felt better even though there was no change in their life circumstances or their attitude toward them (which would be a more authentic improvement).42 We can assume they would report more happiness on a survey, which might lead policymakers to believe incorrectly that citizens are becoming objectively better off (similar to the dis- tortionary effects of adaptation). Another source of happiness that may be regarded as inauthentic derives from relative income, wealth, or status. Psychologists and economists have shown that persons’ relative standing among their peers affects their feelings of happiness and well-being, which Luigino Bruni has termed the “social treadmill,” reflecting the Sisyphean nature of trying to one-up one’s peers while they do the same.43 This can be understood as a psychological drive that benefits society through encouraging competition, but its benefits to individuals in terms of long-term happiness are questionable. Not only does happiness based on relative success fail to increase as everyone works harder, but even the happiness of those who rise above the pack will not be proportionate to their efforts. By implication, if the government were to take an unhappy manager and relocate her among less successful people, her reported happiness may increase—and this may reflect a true increase in felt happiness—but it may not reflect the type of well-being that policymakers want to promote. Even among authentic sources of happiness there are some we may not want to endorse as a society, such as happiness deriving from the bad fortune or sufferings of others. Schadenfreude is a trivial example of this, but there are more distressing cases. Racists and sexists may experience happiness when they see members of a minority group lag behind in some social measure or fail to succeed in general. Even though the majority of society may reject these attitudes and actions based on them, they may be very real sources of happiness to

42. Ronald W. Dworkin, Artificial Happiness: The Dark Side of the New Happy Class (New York: Carroll and Graf, 2006). 43. Luigino Bruni, “Back to Aristotle? Happiness, Eudaemonia, and Relational Goods,” in Capabilities and Happiness, ed. Luigino Bruni, Flavio Comim, and Maurizio Pugno (Oxford: Oxford University Press, 2008), 121–22.

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 18 those who hold them, and if these attitudes are sufficiently prevalent in a community, the effect may be large enough to sway the results of happiness surveys. Furthermore, attitudes toward others’ success do not have to be negative or hateful to be problematic for happiness policy: experiencing happiness when others succeed and prosper may produce “double-counting” in which a policy not only increases happiness among those directly affected but also among others who share their joy but have no direct inter- est in the joyful activity. As philosopher Ronald Dworkin writes (in the language of preferences), “Suppose many cit- “Do we want to izens, who themselves do not swim, prefer the pool to the focus simply on theater because they approve of sports and admire athletes. increasing total . . . Each swimmer will have the benefits of not only his own preference, but also the preference of someone else happiness, or do who takes pleasure in his success.”44 Such socially depen- we want to ensure dent sources of happiness, even benevolent ones that society would generally like to encourage, may confuse policy a more equitable based on happiness measures. distribution as Our discussion of relative effects on happiness brings us back to the goals of happiness-based policy and the issue well?” of distribution. Do we want to focus simply on increasing total happiness, or do we want to ensure a more equitable distribution as well (paralleling the renewed attention given to wealth inequality in recent years)? For instance, consider “utility monsters,” people who derive an extraor- dinary degree of happiness (or misery) from events in their lives. Starting from a fixed set point, such persons (who may be described as particularly sensitive) have a high elasticity of happiness in response to positive or negative events, experiencing great joy from small positive events and tremendous sorrow from minor setbacks. The example of utility monsters serves to highlight the practical differences between maximizing and equal- izing happiness. If a policy were designed that would increase happiness among a certain subset of people, a

44. Ronald M. Dworkin, Taking Rights Seriously (Cambridge, MA: Harvard University Press, 1977), 235.

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 19 maximization goal would indicate that the policy should be focused where it can “earn the greatest return,” namely those who would experience the greatest increase in happiness—the utility monsters—at the cost of the happiness of the less responsive. An equalization goal, on the other hand, would focus on the people who experienced happiness less easily: because the utility monsters’ happiness will jump with the slightest attention, the government can devote most of its resources to increasing the happiness of others to the level of the utility monsters. These problems are familiar to debates over justice and welfare economics, specifically those concerning whether the well-being of the least advantaged is neglected in the single-minded pursuit of maximization.45 These problems cannot be alleviated by simply changing policy targets from GDP or wealth to happiness or subjective well-being, because all such metrics ignore distributional concerns.

CONCLUSION In a 2010 piece on happiness research and policy, philosopher Martha Nussbaum wrote that “the appeal to subjective well-being . . . is riddled with conception confusion and normative naiveté.”46 The problems with happiness- based policy detailed in the current paper are not simply technical or empirical in nature, as maintained by sympathetic researchers all too aware of these issues, and so they cannot be solved simply by developing better processes. These are irresolvable conceptual difficulties of definition, measurement, and implementation, based on the essential nature of happiness as a vague, multifaceted, and subjective phenomenon. Furthermore, policy based on happiness raises a number of ethical and political conundrums that cannot be solved with improved measurement and analysis, because answers to these questions lie in the realm of philosophy, not science. Setting all the conceptual difficulties aside, there remains the issue of why happiness should be an important concern of government policy at all. Government can be assigned certain roles and responsibilities depending on a person’s political orientation, ranging from the provision of essential public services to alleviating suffering among the needy and disadvantaged. But it is hard to imagine that many liberals, conservatives, or libertarians would endorse the promotion of aggregate happiness as a priority of government. Each ideology

45. Examples include John Rawls’s difference principle or Derek Parfit’s prioritarianism, both of which focus on benefitting the least well-off. See Rawls, A Theory of Justice (Cambridge, MA: Harvard University Press, 1971); Parfit, “Equality and Priority,” Ratio 10, no. 3 (2002): 202–21. 46. Nussbaum, “Who Is the Happy Warrior?,” 108.

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 20 has different ideas about the scale and scope of government, but they all share a focus on certain services or resources that the government should provide to the people to enable them to pursue their own idea of happiness. If we cannot justify policymakers’ focus on happiness, support for the continued development and use of these fundamentally flawed measures is questionable at best. In his book The Politics of Happiness, former Harvard president Derek Bok details many of the same reservations presented above, but nonetheless concludes that “in the end, the relevant question in making policy is not whether self-evaluations of well-being are perfect, but whether they are at least as accurate as the best alternative ways of gauging people’s preferences, opinions, and needs.”47 The suggestion that current happiness measures are “good enough for now” is hard to accept without also demonstrating the need for devoting so many resources to studying something as ephemeral as happiness or subjective well-being in the first place. Instead, these conceptual problems suggest that the government should not be trying to study or influence happiness directly at all, even as one measure among many. The conceptual difficulties with happiness policy only get more dire once it is acknowledged that we cannot know the true nature of any one person’s happiness, well-being, or interests in general, much less that of an entire population. Given the complex, multifaceted, and subjective nature of happiness, any measure of it designed by researchers and policymakers is guaranteed to impose their conception of happiness on the subjects of the study rather than measure the subjects’ happiness according to their own conceptions of it (as imperfect as those may be). For this reason, Sissela Bok describes definitions of happiness as Rorschach tests that reveal much about the preconceptions and values of those proposing the definitions.48 Even if policymakers have the best of intentions, they cannot help but substitute their idea of what makes people happy for what people actually care about. This defeats the stated purpose of happiness-based policy and its oft-cited advantage over traditional measures such as GDP. Happiness-based policy shares this feature with another relatively new policy approach, libertarian paternalism, in which policymakers exploit natural cognitive biases and heuristics to steer or “nudge” people into choices that they would have made themselves, were their choices not influenced by those very same cognitive flaws.49 Both happiness policy and libertarian paternalism draw

47. Derek Bok, The Politics of Happiness: What Government Can Learn from the New Research on Well-Being (Princeton, NJ: Princeton University Press, 2010), 39. 48. Sissela Bok, Exploring Happiness, 54–58. 49. Richard H. Thaler and Cass R. Sunstein, Nudge: Improving Decisions about Health, Wealth, and Happiness (New Haven, CT: Yale University Press, 2008).

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 21 from new psychological research in an attempt to craft policy with the intent of improving citizens’ well-being, and both also share an unjustified presumption of knowledge about what makes up people’s conceptions of what makes them better off.50 Policies that nudge people in the direction of saving more for retire- ment or eating less fast food assume that savings or dietary health are the only interests that people have while making those choices, when they might have any number of reasons for making them. By the same token, policies based on arbitrarily chosen and ambiguously measured conceptions of happiness have little chance of capturing what actually makes people happy or better off. The psychological science underlying both policy approaches may be valid, but policymakers lack the necessary information to use them to increase people’s actual happiness or well-being. Instead, policymakers end up promoting whatever conceptions of happiness or well-being they think people have—or that they think people should have. In this age of “big data,” increasing acceptance of personal surveillance, and decreasing respect for the individual’s autonomy, policy approaches such as these have enormous appeal to the government and seem innocuous to much of the citizenry. While “nudging” people into making choices for their own good has encountered objections based on its explicit paternalism, happiness- based policy might be immune, despite similar concerns, because the policy tools are more blunt and less focused—and increasing happiness rather than economic output is likely to have significant appeal to voters (for good reason). Furthermore, if concerns about inequality continue to grow, so will skepticism about the relevance of economic measures such as GDP to the average person on the street, and happiness may seem like a more egalitarian focus for policy (although, as described above, it is not). If we acknowledge that well-being makes a better target for government policy than economic output or GDP—which is reasonable—then the shortcomings of happiness measures introduce a challenge: how to promote the actual well-being of citizens if the government cannot accurately and reliably measure it. One suggestion would be to improve the measurement of GDP itself to incorporate many of the elements of well-being that it cur- rently ignores, such as nonmarket output, externalities, and distribution. It would still remain primarily a measure of output, but it would more closely approximate a measure of the resources available to the average citizen to pursue his or her own conception of well-being, without getting mired in the

50. For more on this line of argument against nudge, see Mark D. White, The Manipulation of Choice: Ethics and Libertarian Paternalism (New York: Palgrave Macmillan, 2013).

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 22 issues of subjectivity, adaptation, and interpretation that plague the measurement of happiness. A related suggestion, mentioned above, is to shift policymaking emphasis away from directly targeting economic output and well-being and focus it on structuring institutions to allow people the greatest amount of personal choice possible (consistent with all others having the same) to pursue their own goals and interests. This would free up government resources to deal with ongoing problems and sudden crises when they present themselves, rather than trying to manage the economy by finding problems in artificially created data based on arbitrary definitions and ambiguous measurements. Our elected leaders could promote aggregate well-being more successfully if they concentrated their attention on specific problems of national importance and left the pursuit of happiness to those best qualified for it: the people.51

51. For more detail on this policymaking idea (as well as the earlier critiques of the measurement and policy use of happiness and well-being), see Mark D. White, The Illusion of Well-Being: Economic Policymaking Based on Respect and Responsiveness (New York: Palgrave Macmillan, 2014).

MERCATUS CENTER AT GEORGE MASON UNIVERSITY 23 ABOUT THE AUTHOR Mark D. White is chair and professor in the Department of Philosophy at the City University of New York’s College of Staten Island, where he teaches courses in philosophy, economics, and law. He has published widely in the intersections of these areas, including four books, more than 40 journal arti- cles and book chapters, and over a dozen edited volumes. Two of his recent books, The Manipulation of Choice (2013) and The Illusion of Well-Being (2014), identify problems with emerging trends in policymaking such as nudges and happiness measures. He recommends instead orienting policymaking according to respect for individuals and their interests. ABOUT THE MERCATUS CENTER AT GEORGE MASON UNIVERSITY The Mercatus Center at George Mason University is the world’s premier university source for market-oriented ideas—bridging the gap between academic ideas and real-world problems. A university-based research center, Mercatus advances knowledge about how markets work to improve people’s lives by training graduate students, con- ducting research, and applying economics to offer solutions to society’s most pressing problems. Our mission is to generate knowledge and understanding of the institutions that affect the freedom to prosper and to find sustainable solutions that overcome the barriers preventing individuals from living free, prosperous, and peaceful lives. Founded in 1980, the Mercatus Center is located on George Mason University’s Arlington campus.